Files
pytorch/caffe2
Nicolas De Carli 2459da4a64 [Caffe2] Add float batch box cox SVE128 implementation (#159778)
Introduce SVE128 SIMD batch box-cox computation.

We've seen about 65% throughput improvement.

Privacy Context Container: L1196524

This is a no-op from OSS point of view, therefore it could be landed without tests (see precedence set by https://github.com/pytorch/pytorch/pull/143627), but we should delete those at some point

Pull Request resolved: https://github.com/pytorch/pytorch/pull/159778
Approved by: https://github.com/malfet
2025-09-16 07:25:04 +00:00
..